Affiliations: [a] CSIRO Health and Biosecurity, Brisbane, Australia
| [b] Department of Aged Care, Austin Health, Heidelberg, Victoria, Australia
| [c]
University of Melbourne, Parkville, Victoria, Australia
| [d]
Melbourne Dementia Research Centre, Parkville, Victoria, Australia
| [e]
Florey Institute of Neuroscience and Mental Health, Parkville, Victoria, Australia
| [f] CSIRO Data61, Brisbane, Australia
Correspondence:
[*]
Correspondence to: Scott Ayton, The Melbourne Dementia Research Centre, The Florey Institute of Neuroscience and Mental Health, 30 Royal Parade, Parkville, Victoria 3052, Australia. Tel.: +61 3 9035 6559; E-mail: scott.ayton@florey.edu.au.
Note: [1] These authors contributed equally to this work.
Abstract: Background:Cortical iron accumulation has been reported as a pathological feature of Alzheimer’s disease (AD). The cause of cortical iron elevation in AD is unknown but may be contributed by hemosiderin deposits in cerebral microbleeds that frequently occur in this disease. Objective:To investigate the impact of cerebral microbleeds (which are more frequent in AD) on the magnetic susceptibility of the surrounding brain tissue. Methods:32 MRI scans from the Australian Imaging, Biomarker and Lifestyle (AIBL) study were found to have cerebral microbleeds by manual assessment of susceptibility weighted images. Quantitative susceptibility mapping (QSM; an MRI technique that is sensitive to iron) was used to estimate iron content in the tissue surrounding the microbleed in four concentric radii. Furthermore, the mirror regions on the contralateral hemisphere were also demarcated. A simulation analysis was conducted to investigate the effect of QSM imaging on cerebral microbleeds with varying sizes. Results:77 microbleeds were identified from the available scans. The immediate proximal region to the cerebral microbleeds had enhanced tissue susceptibility (∼0.02 PPM), but importantly, this did not extend beyond one voxel radius. This finding with in vivo data was also replicated in a simulation study. However, the presence of microbleeds could lead to over-estimation of tissue QSM in unsupervised quantification, therefore processing methods to avoid this artefact without the need for their manual identification are proposed. Conclusion:The local changes in susceptibility due to microbleeds outside the focal lesion are restricted to 1 voxel and may be explained by partial voluming artefacts caused by limited imaging resolution. The susceptibly change induced by the microbleed is a relatively small proportion of tissue and could not account for regional iron changes observed in AD cortex.
